Why Don’t DP Vessels Measure Thrust Better?

Introduction:?Pitch and speed feedbacks are sometimes poor estimates of thruster force, but they are the primary means that most DP control systems use to estimate force.?DP incidents have been caused by a large difference between the estimated thruster force and the actual force.?It’s such a large error, Kongsberg has used it as an argument against measuring water current, as the errors in the thrust measurement often dwarf the effect of water current on the vessel.?So, why don’t we measure it better?

What’s Wrong with Pitch or Speed??They are indirect thrust measurements.?Assuming everything stays in the same shape and conditions, thrust is roughly proportional to pitch in the linear portion of the thrust curve (constant speed), and thrust is proportional to the square of the speed (constant pitch).?Variable pitch and speed can then be calculated for combination control in accordance with the propeller and control system characteristics.?But everything doesn’t stay the same and conditions change.?Errors in speed, or pitch, measurement or control throw these calculations off, and conditions change regularly.?Water flow from other thrusters, the ocean current, & vessel motion, and water conditions, such as air bubbles, temperature, contaminants, & objects, can have a considerable effect on thrust output.?If only there were a way to measure thruster force output independent of these errors, DP control could be improved.

Torque/Force Control:?There is, and it has been used for years.?A high input current (flow into thruster suction) reduces thrust and thruster load for a given speed and pitch.?Torque feedback is less popular than speed and pitch feedback, but has earned a good reputation among experienced adherents, and is supported by most major DP control system providers.?Usually torque is measured at the motor or engine driving the thruster, so it includes mechanical losses, as well as thrust force.?As these losses can also be variable (e.g. bearing, lubrication, gear, and coupling faults), there were even thrusters with load cells built into the propeller hubs to measure propeller force directly.?These were all improvements, but they were not perfect solutions.?Errors in measurement are still a potential problem, and the hub load cells were subject to considerable vibration and very difficult to maintain.?Even if the propeller force is measured correctly, that still doesn’t mean that the net thrust force is measured correctly, as the thrusters interact with the hull and things attached to it.

Horror Story:?Hull interaction isn’t much of a problem on large, semi-submersables and barges, with flat bottoms and azimuth thrusters angled down 8° to avoid the thruster wash interacting with the hull, but it can be an enormous problem on regular ship-shaped vessels.?When I was new to DP, I had a nasty early experience on a new-build vessel with offset rudders.?The rudders were offset to allow the main propeller shafts to be removed for maintenance without having to remove the rudders.?This meant that each was more effective in one rudder direction than in the other, but the DP control system provider refused to model this, and depended on the DP model to compensate for unbelievable errors.?Rudder swings or change of active rudder created huge changes in DP current (the sum of all unknowns), and the vessel could not hold position when the active rudder or main prop was lost, as the DP model had built up too large a body of lies to compensate for the incorrect tuning.?If it had been a major drill company or vessel owner, the DP company would have fixed its base modeling to adapt to the vessel (seen it many times), but as it was a smaller company at the time, the vessel’s mechanical construction needed to be changed to make the DP system work (true story).?It wouldn’t matter how well propeller thrust was measured, as the interaction with the rudders was not modelled properly in the DP system, and systems with large false DP currents cannot adapt to sudden change.?This is a cautionary story, and an extreme case of failure to properly reflect the rudder design, but many DP vessels fail to model considerable hull interactions with their thrusters.?If a leading DP company wouldn’t bother modeling something as obvious as offset rudders, what chance do more subtle threats have?

Hull Interaction:?Ignoring rudders and the associated horror story, there are a lot of things that can reduce the effective thrust from the propeller.?We can measure the propeller force fairly well (assuming no damage to the propeller and good calibration), but that is not the same as measuring net thrust, as the losses are not measured and often not directly modeled.?For example, a thruster whose wash drags across the hull is exerting a force on the hull, and effective net thrust is reduced by this counter force.?It’s a bit like biasing two thrusters against each other, the net bias does not add to overall vessel propulsion force, but does burn more fuel.?These hull loses are directional and variable depending on conditions.?For example, main azimuth thrusters have more thrust lost from wash over the forward hull, against another thruster, and especially against the skeg, than into empty water.?Similarly, vessel motion, thruster wash interaction, and water current interaction can increase or decrease the application of these forces.?This effect of changing conditions cannot be directly modeled.?Ideally, the first, structural part would be carefully measured and modeled during DP scaling.?But DP scaling does not occur in repeatable laboratory conditions, and owners are always in rush to get the vessel out to work, so the vessels are roughly scaled and the DP model used to compensate for unknowns.?Net thrust is the important thing that the DP system really wants to know, but cannot directly measure due to practical and commercial limitations.

Calibration:?The main problem with measuring anything is having an effective means of measurement and maintaining the accuracy of that measurement.?Experienced marine engineers are aware of the constant battle to keep equipment within specifications within the allowed time and cost.?DP control system providers have seen enough of this fight, and its occasional results, that they are more comfortable depending on the DP model to correct problems, than they are depending on the vessel owners to maintain thrust measurement calibration.?They have a point.?Of course, the more accurate the information and base model, the lower the DP current component of those errors, and the better the vessel is at adapting to sudden changes without those false assumptions.?Accuracy costs money, and many owners want to limit how much they invest in it.??

Go Slow:?This is a major reason why DP vessels are only allowed to make small and gradual changes during critical operations.?Too large a change, when there is a large difference between reality and the DP model, can cause ill adapted control decisions that cause loss of position.?The difference between the DP model and reality isn’t always known, but should be estimated by comparing the DP and water current.?Slow and steady keeps position, but faults can cause sudden change, so errors need to be kept within acceptable limits.?What is acceptable??It varies from vessel to vessel and depends on practical testing, but if the DP current follows changes in the ship’s heading, or if it makes large changes if different thrusters are used, then redundancy is endangered.

Conclusion:?The difference between normal and good calibration is large enough that the difference between torque thrust control and speed, or pitch, is usually less vital.?Better control with less errors is available to those willing to invest in it and maintain it, but the general market is less interested, and happy making slow, small heading & position changes, to avoid trouble and costly investments.?The market is happy with thrust measurement that is “good enough”, and the market’s low standards keep pressure off of the DP control system suppliers for improvement.?DP control system improvement cannot occur until owners take responsibility.?In the meantime, operators need to keep a careful eye on the system’s behavior, and vessel clients need to give time for calibration and maintenance.?We can do better, but it costs time and money, and pitch and speed are good enough for most applications.